Memory module prices have plummeted, with 16GB modules dropping over RMB 100 in a single day, driven by large-scale selling from major stockpilers. The decline follows a significant drop in market sales, which have decreased by more than 60% since November last year due to previously inflated prices and insufficient demand. Contributing to the market downturn, Google introduced "TurboQuant," a new compression algorithm that reduces key-value cache memory usage by at least 60% during the runtime of large language models. This innovation has led to lowered expectations for AI hardware demand, impacting U.S. memory chip stocks, which lost nearly $100 billion in market value this week. Micron Technology's stock alone fell 15%, erasing over $70 billion in market value.